Xenserver Space Woes
-
@scottalanmiller said in Xenserver Space Woes:
@dafyre said in Xenserver Space Woes:
It looks as if his XenServer is reporting that the VMs are using 2x the space that they should be.
Can we determine if it is exactly 200%? If so... maybe the SAN is doing some kind of internal replication?
@jrc says... Xen Center reports 7378.5Gb Used, 4147.8Gb allocated out of 11166.3Gb
So, not quite 2x. But right close.
-
@dafyre said in Xenserver Space Woes:
@scottalanmiller said in Xenserver Space Woes:
@dafyre said in Xenserver Space Woes:
It looks as if his XenServer is reporting that the VMs are using 2x the space that they should be.
Can we determine if it is exactly 200%? If so... maybe the SAN is doing some kind of internal replication?
@jrc says... Xen Center reports 7378.5Gb Used, 4147.8Gb allocated out of 11166.3Gb
So, not quite 2x. But right close.
That's almost 300GB less than double the allocated. It's not the SAN doing something.
What kind of retention / backup system do you have in place?
-
@scottalanmiller said in Xenserver Space Woes:
@dafyre said in Xenserver Space Woes:
It looks as if his XenServer is reporting that the VMs are using 2x the space that they should be.
Can we determine if it is exactly 200%? If so... maybe the SAN is doing some kind of internal replication?
I am nearly 100% sure the SAN is not doing anything more than presenting it's space to Xen. The fact that it's close to double the allocated is, I think, merely a co-incidence based on the fact that one of the VMs has a 2Tb VDI assigned to it that is mostly full. So I am pretty sure the problem is some sort of snapshot issue.
When I run xe vdi-list is-a-snapshot=true I get:
uuid ( RO) : 53d7329c-00fe-4659-a9fa-779e6341637e
name-label ( RW): Staff Home
name-description ( RW): VDI for staff home folders
sr-uuid ( RO): 4558cecd-d90d-3259-7ea5-09478d0e386c
virtual-size ( RO): 2193654546432
sharable ( RO): false
read-only ( RO): trueuuid ( RO) : 586f6e9a-bf7d-4fc4-89aa-6568ba91cea5
name-label ( RW): Staff Home
name-description ( RW): VDI for staff home folders
sr-uuid ( RO): 4558cecd-d90d-3259-7ea5-09478d0e386c
virtual-size ( RO): 2193654546432
sharable ( RO): false
read-only ( RO): trueNotice that both seem to be snapshots, and each is roughly 2Tb in size. So I think these are my culprits. Though what I can do with them I have no idea...
-
@DustinB3403 said in Xenserver Space Woes:
What kind of retention / backup system do you have in place?
A non-working Unitrends implementation (another freaking thing I am trying to solve, and am starting to think it may be related to this problem).
-
@jrc said in Xenserver Space Woes:
@scottalanmiller said in Xenserver Space Woes:
@dafyre said in Xenserver Space Woes:
It looks as if his XenServer is reporting that the VMs are using 2x the space that they should be.
Can we determine if it is exactly 200%? If so... maybe the SAN is doing some kind of internal replication?
I am nearly 100% sure the SAN is not doing anything more than presenting it's space to Xen. The fact that it's close to double the allocated is, I think, merely a co-incidence based on the fact that one of the VMs has a 2Tb VDI assigned to it that is mostly full. So I am pretty sure the problem is some sort of snapshot issue.
When I run xe vdi-list is-a-snapshot=true I get:
uuid ( RO) : 53d7329c-00fe-4659-a9fa-779e6341637e
name-label ( RW): Staff Home
name-description ( RW): VDI for staff home folders
sr-uuid ( RO): 4558cecd-d90d-3259-7ea5-09478d0e386c
virtual-size ( RO): 2193654546432
sharable ( RO): false
read-only ( RO): trueuuid ( RO) : 586f6e9a-bf7d-4fc4-89aa-6568ba91cea5
name-label ( RW): Staff Home
name-description ( RW): VDI for staff home folders
sr-uuid ( RO): 4558cecd-d90d-3259-7ea5-09478d0e386c
virtual-size ( RO): 2193654546432
sharable ( RO): false
read-only ( RO): trueNotice that both seem to be snapshots, and each is roughly 2Tb in size. So I think these are my culprits. Though what I can do with them I have no idea...
If this is a snapshot issue, and you have no snapshots shown on the VM >Snapshots tab, for this VM I'd rescan your storage, and see if it is cleaned up.
Otherwise you'll have to reboot the system, let the system scan the storage.
-
Are you planning to stick with Unitrends and just fix it?
-
@DustinB3403 said in Xenserver Space Woes:
Are you planning to stick with Unitrends and just fix it?
Well at this point if Unitrends would start to work, then yes we'll stick with it, at least until the support contract we paid for is up, at which point we'll go back out and see what other options are available to us. If I can't get Unitrends to work without having to re-write the code myself, then we may end up ditching it.
@DustinB3403 said in Xenserver Space Woes:
If this is a snapshot issue, and you have no snapshots shown on the VM >Snapshots tab, for this VM I'd rescan your storage, and see if it is cleaned up.
Otherwise you'll have to reboot the system, let the system scan the storage.
There are no snapshots listed in the VM's snapshot tab, and when I rescan the storage, nothing seems to happen or at least nothing changes. And as I said, I'll be rebooting the f-ing things this weekend.
-
@jrc said in Xenserver Space Woes:
@DustinB3403 said in Xenserver Space Woes:
Are you planning to stick with Unitrends and just fix it?
Well at this point if Unitrends would start to work, then yes we'll stick with it, at least until the support contract we paid for is up, at which point we'll go back out and see what other options are available to us. If I can't get Unitrends to work without having to re-write the code myself, then we may end up ditching it.
@DustinB3403 said in Xenserver Space Woes:
If this is a snapshot issue, and you have no snapshots shown on the VM >Snapshots tab, for this VM I'd rescan your storage, and see if it is cleaned up.
Otherwise you'll have to reboot the system, let the system scan the storage.
There are no snapshots listed in the VM's snapshot tab, and when I rescan the storage, nothing seems to happen or at least nothing changes. And as I said, I'll be rebooting the f-ing things this weekend.
~ Reboot now! No time for the weekend.. If it's gonna break its gonna break on company time damn it! ~
-
@DustinB3403 said in Xenserver Space Woes:
~ Reboot now! No time for the weekend.. If it's gonna break its gonna break on company time damn it! ~
Well it's not broken now, everything is working as expected. So if I can schedule the down time it'll be better for everyone. However I am getting a "Run out of space while coalescing," which come to think on it may also be related...
-
@jrc said in Xenserver Space Woes:
@DustinB3403 said in Xenserver Space Woes:
~ Reboot now! No time for the weekend.. If it's gonna break its gonna break on company time damn it! ~
Well it's not broken now, everything is working as expected. So if I can schedule the down time it'll be better for everyone. However I am getting a "Run out of space while coalescing," which come to think on it may also be related...
Oh man... You just messed up. Now it is definitely going to break!
-
@dafyre said in Xenserver Space Woes:
@jrc said in Xenserver Space Woes:
@DustinB3403 said in Xenserver Space Woes:
~ Reboot now! No time for the weekend.. If it's gonna break its gonna break on company time damn it! ~
Well it's not broken now, everything is working as expected. So if I can schedule the down time it'll be better for everyone. However I am getting a "Run out of space while coalescing," which come to think on it may also be related...
Oh man... You just messed up. Now it is definitely going to break!
Well shit.... looks widely around for some wood to know on
-
Make sure you have a backup before you reboot.
-
@brianlittlejohn said in Xenserver Space Woes:
Make sure you have a backup before you reboot.
Well so you see there's the rub. I suspect my backups aren't working because of this issue, and the potential fix is to reboot. So I am damned if I do, and I am damned if I don't.
-
If you goto the Storage in your Xencenter (i assume yorue using xencenter to look at things) Go to the SR with trouble, and look at the Storage tab for that SR. In the disks section you may see those 2 'orphaned' vdi, as well as the real one.
Is that the case? are they listed in Xencenter>SR>Storage? -
@momurda said in Xenserver Space Woes:
If you goto the Storage in your Xencenter (i assume yorue using xencenter to look at things) Go to the SR with trouble, and look at the Storage tab for that SR. In the disks section you may see those 2 'orphaned' vdi, as well as the real one.
Is that the case? are they listed in Xencenter>SR>Storage?The are not listed there. The only VDIs listed there are the ones actually being used by all my VMs.
-
Ok, what about the SR where you Unitrends backup disks are? Is there anything amiss there?
-
@momurda said in Xenserver Space Woes:
Ok, what about the SR where you Unitrends backup disks are? Is there anything amiss there?
Nope. Shows only the backup target VDIs (4 of them, each 2044Gb)
-
Hmmm. I ran xe vdi-list is-a-snapshot=true
Got 1 result, a disk from our crm server. If you go to
the Objects button instead of Infrastructure in XC(
You see in mine here, the 2 disk for crm01. One of them shows as a snapshot, even though the XC thinks there isnt one for this vm. This is a byproduct of failed backups i think with Unitrends, soemtimes they dont delete themselves.
If you do this you may see your 2 vdi listed here as well as the disk actually being used.
Ah weird formatting on the pic, soz. -
I don't have a Virtual Disks option that I can find, certainly nothing that looks like what you have in the tree. How did you get there? How did you get it to show up like that?
NM, that's in the Object View. And no, the two disks I mentioned earlier do not show up in that list either.
-
This is the Objects view, not Infrastructure view in Xencenter.
From there expand Virtual Disks, and you will see this view.
Edit: Have you selected View Hidden Objects from View menu?
The problem seems to be these 2 massive disks that are there taking up space, but not being seen by XC correctly. I am certain you could manually delete them, but i would make sure the 'real' disk that you are currently using has a different uuid than the 2 that are in limbo.
